Stay Organized at Scale: Introducing Prompt Groups
Are you a Conductor customer?
Review this article—and the rest of our educational content and activities—in our in-platform Community!
Not a customer? Click here to view this article
🔒 Early Access Feature: Prompt Groups is currently available in alpha to select Conductor customers. If you don't see this feature in your account yet, it's on its way — reach out to your Customer Success Manager to learn more about early access.
If you've been using AI Search Performance for a while, there's a good chance your prompt library has grown — a lot. Topics gave you a solid thematic framework. But as teams track hundreds of prompts across multiple initiatives, campaigns, and business priorities, one question keeps coming up: "How do I cut through all of this and focus on the prompts that actually matter to what I'm working on right now?"
Prompt Groups is the answer. It's a new way to label, organize, and analyze your tracked prompts by any dimension that matters to your team — not just topic. A campaign. A product launch. A business unit. A set of prompts your team added manually versus ones Conductor's AI generated. Whatever lens you need, you can now build a group around it.
In this guide:
Groups vs. Topics — What's the Difference?
How Prompt Groups Work
Ways to Create a Group
Rule-Based Groups: The Smart Way to Organize at Scale
Use Cases: What Can You Group By?
Locking a Group
Putting It All Together
Before diving in, it's worth getting clear on how Groups and Topics relate — because they're complementary, not competing.
Topics are thematic. They're organized around the subject matter of the prompt — "hybrid mattress," "mattress buying guide," "best mattress for back pain." Topics are the semantic anchor of your tracking setup and drive how content clusters in the report.
Prompt Groups are organizational. They don't need to follow any thematic logic. A group can pull from multiple topics, go narrower than a single topic, or cut across your entire prompt library using a completely different dimension — like which team set up the prompts, which campaign they support, or which AI engine context they apply to.
Think of Topics as what you're tracking. Groups are how you want to work with it.
Every group you create has a rule that defines which prompts belong to it. That rule can be as simple as "any prompt I manually add to this group" — or as sophisticated as "all branded prompts, tracked on ChatGPT, assigned to the Consideration intent, that were added after January 1st."
Group membership is automatically maintained. When the conditions in your rule are met by a new or updated prompt, it joins the group. When they no longer apply, it leaves. You don't have to manually update groups as your prompt library evolves.
You can also override the rule at any time by explicitly including or excluding specific prompts — giving you the precision of a rule-based system with the flexibility of manual control.
There are two approaches, and they can be combined.
Manual Groups
You pick the prompts. You build the list. The group stays exactly as you define it. Best for curated sets that don't need to automatically expand — like a fixed list of high-priority prompts for an upcoming board review, or a specific set of prompts a colleague has asked you to monitor.
Rule-Based Groups
You define one or more conditions, and Conductor automatically populates the group with every prompt that matches — now and in the future. Best for anything that will grow or change over time, or where you want consistent, repeatable reporting without ongoing manual maintenance.
The two approaches can coexist in a single group. You can set a rule that pulls in prompts automatically, then pin specific prompts that should always be included regardless of the rule — or exclude specific ones that should never appear even if they technically match.
Pro Tip: Start with a rule-based group and use manual inclusions sparingly for exceptions. This way your group stays accurate and self-updating as your prompt library grows, without requiring regular maintenance.
When you create a rule-based group, you're building a filter that runs continuously against your tracked prompts. You can combine conditions using AND (a prompt must match all conditions in a criteria set) and OR (a prompt can match any one of multiple criteria sets).
Here's what you can filter on:
You can stack these filters to build highly targeted groups. A few examples of what that looks like in practice:
All unbranded prompts, on ChatGPT, with Purchase intent → a group for your bottom-of-funnel conversion tracking
All AI-generated prompts, created after your last configuration session → a group for reviewing and auditing Conductor's auto-generated prompts before accepting them
All prompts belonging to Topic A or Topic B, assigned to the "Enterprise Buyer" persona → a cross-topic group scoped to a specific audience segment
The most useful grouping dimensions depend on how your team works. Here are some of the most common and highest-value ways to use Prompt Groups.
By Campaign or Initiative
Track performance for a specific campaign without mixing it in with your broader reporting. Create a group for "Spring Product Launch" that contains all the prompts relevant to that initiative — pulling from multiple topics if needed. When the campaign ends, the group is still there for retrospective analysis.
By Business Line or Product Category
If your company operates across multiple product lines or business units, groups let each team work with their own slice of the data. A "Consumer Audio" group and a "Professional Audio" group can coexist in the same account, each scoped to the topics and prompts that matter to those teams respectively.
By AI Engine
Want to understand how your brand performs on Perplexity specifically, compared to ChatGPT? Create engine-scoped groups and analyze them side by side. This is especially useful when different engines are treating your brand very differently.
By Who Set Up the Prompts
Conductor's AI can auto-generate prompts based on your topics. Your team can also add prompts manually. With Prompt Groups, you can filter by source — separating AI-generated prompts from manually created ones — so you can review, audit, or analyze each set independently.
Use case: Create a group containing only AI-generated prompts to do a regular quality check. Review performance, decide which ones to keep tracking, and remove anything that isn't relevant — without having to scroll through your entire prompt library to find them.
By Priority or Business Importance
Not all prompts are created equal. Some represent your highest-value search conversations — the ones your CMO cares about, the ones tied to revenue-driving topics. Create a "High Priority" or "C-Suite Watch List" group and populate it with only those prompts. Use it as the basis for your executive reporting so leadership always sees the most strategically important data, not a noisy average across everything.
As a "Sub-Topic" View
Sometimes a single topic is too broad for the analysis you need. "Mattress buying guides" might contain 200 prompts — but you only care about the ones that include a specific keyword phrase or are assigned to a specific persona. A group lets you create a granular sub-topic view without changing your topic structure.
As a "Cross-Topic" View
Some initiatives span multiple topics. A "Brand Awareness Campaign" might have relevant prompts across "product comparisons," "buying guides," and "brand alternatives" topics. A group lets you pull all of those together into one view, even though they don't share a topic.
Once a group is set up and becomes a load-bearing part of your reporting — a "Brand Keywords" group that feeds your weekly executive dashboard, for example — you probably don't want it changed accidentally.
Locking a group protects it from edits. While locked, no one can rename it, modify its rule, add or remove prompts from it, or delete it. The group remains fully usable in reports, dashboards, and analysis — it just can't be touched.
Only the group's creator or an account admin can lock or unlock a group.
Pro Tip: Lock any group that's connected to a scheduled report, a Workspace, or a reporting cadence you rely on. It takes two seconds and prevents the kind of accidental edit that only gets noticed three weeks later when a dashboard looks wrong.
Prompt Groups work best when you think of them as a layer of operational organization on top of your semantic topic structure. Topics tell the story of what your brand should own in AI search. Groups make your daily analysis workflows faster, your team reporting cleaner, and your large prompt libraries actually manageable.
A few principles worth keeping in mind as you build:
Use rule-based groups wherever possible. They stay accurate automatically. Manual groups need maintenance.
Don't over-group. Start with the dimensions that map directly to how your team reports and works. A "Q3 Campaign," "AI-Generated Prompts Audit," and "Executive Watch List" are more useful than twenty micro-groups.
Lock anything connected to recurring reporting. It's a simple safeguard that prevents downstream dashboard problems.
Use groups to scope team reporting. If different people on your team own different areas, give each person their own group. They get a clean view of what they're responsible for, without noise from the rest of the account.
Ready to build your first group? If you have access to the alpha, head to Settings > Topics and Prompts and look for the Groups option in your prompt management settings. Not seeing it yet? Prompt Groups is currently rolling out to select customers — contact your Customer Success Manager to find out about access.